I am going through the set of features available in lf and resolving all of the ones that are marked as experimental. It turns out that locale is the only experimental feature, which I must decide whether to accept it in full (removing its experimental status), or to reject it completely.

While I can respect the need for users to sort based on locale, I have decided on this occasion to reject this feature, for the following reasons:

• It is not a common requirement. File managers typically sort based on only file stat fields such as name/mtime/size, although exceptions are made for cases like natural sorting and sorting by extension as they are frequently required by users.
• The feature requires go-locale as a dependency. While there is likely nothing wrong with go-locale itself, in general having more dependencies means more risk, and I am hesitant to keep this for a feature that I don't consider as widely used.
• The text/collate package is updated once in a while, and I have to briefly test changes before merging Dependabot bumps. It is not a big deal, but still requires some effort on my part.
• Since the addition of custom sorting (see Sort by custom method #618 for more details), this can be implemented via user configuration. Continuing to include locale sorting as a built-in feature goes against the minimalistic design goals of lf (i.e. make features simple but extensible, rather than trying to patch in direct support for everything).
• Furthermore, if custom sorting had been implemented first, I would not have merged this feature in the first place. The only reason locale sorting still exists now is simply just because it came first.